Resection Margins

Radical surgery with curative intent is the treatment of choice in the majority of colon cancers. The basic surgical principles are removal of the major vascular pedicle feeding the tumor along with its lymphatics, obtaining a tumor-free margin, and en bloc resection of any organs or structures attached to the tumor. True colonic mucosal recurrences are rare. More common are para-anastomotic recurrences reflecting possibly an inadequate lymphadenectomy. It is therefore recommended that at least a 5-cm margin of normal bowel be obtained on either side of the tumor in order to minimize the possibility of an anastomotic recurrence.²⁵⁴ For right-sided tumors, the length of ileum apparently does not influence the local recurrence rate.²⁵⁴ Figures 106-2 and 106-3 note the extent of resection. It is important to note that for tumors in the cecum, ascending colon, hepatic flexure, and proximal transverse colon, the right branch of the middle colic artery is divided along with the right colic and the ileocolic arteries. If the middle colic artery is ligated at its origin, consideration should be given to extend the resection of the bowel just to the distal third of the transverse colon in order to ensure viable bowel for the anastomosis. Tumors in the transverse colon may require transverse colectomy or, at times, an extended right colectomy where the cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, and upper descending colon are resected together with its lymphatic drainage. Descending and upper sigmoid colon cancers can be treated with left hemicolectomy with ligation of the IMA at its origin from the aorta or by segmental resections as long as the principles outlined above are followed.

Figure 106-2

Segments of bowel and lymph node containing mesentery to be removed for carcinoma of the cecum (A-A'), hepatic flexure (A-B), splenic flexure (C-C'), and descending colon (C-D).

Figure 106-3

Segments of bowel and lymph node containing mesentery to be removed for carcinoma of the transverse colon, the apex of the sigmoid (A-B), and the lower sigmoid or rectosigmoid (A-A').

All these procedures can be performed with a hand-sewn or stapled anastomosis. It is of utmost importance that there is no tension at the anastomosis otherwise it will fail with resultant leak and sepsis.

Lymphadenectomy

Because there are patients with stage III colon cancer that are cured by surgery alone, performance of an adequate lymphadenectomy cannot be overemphasized. Adequate lymph node resection is imperative for adequate staging and selection of patients for adjuvant treatment. Whether a wide resection or an extended resection is necessary is still controversial. If possible, disease outside the resection field should be documented, as it will have an impact on prognosis, as well as on eligibility for any clinical trials. A minimum of 12 negative lymph nodes should be examined to accurately define node negative disease.²⁵⁴

Lymphatic Mapping

Lymph node metastases in colorectal cancer appear to be an orderly process. Lymphatic mapping identifies the lymph node(s) that has(ve) the highest probability of harboring metastatic disease. By identifying the sentinel lymph node, immunohistochemical techniques and molecular techniques can be used to evaluate for the presence of micrometastases otherwise not diagnosed by conventional pathologic examination. In both single-institution and multiinstitutional studies it has been shown that lymphatic mapping is feasible and indeed may upstage 14% to 18% of node-negative tumors to node-positive status.^255,256 The significance of lymph node micrometastases in colorectal carcinoma is unknown. Similar to reports of micrometastases in the bone marrow of patients with colorectal carcinoma undergoing curative resection, a report from The Netherlands suggests that lymphatic micrometastases adversely affect prognosis.^257,258 Prospective studies are needed to clarify the significance of molecular detection of lymphatic micrometastases.

Obstruction and Perforation

Colorectal cancer presenting with obstruction or perforation has an adverse effect on survival. In two prospective randomized trials of Dukes Band C colorectal trials, the National Surgical Adjuvant Breast and Bowel Project (NSABP) reported that patients with bowel obstruction were at greater risk for treatment failure than were those patients without obstruction.²⁵⁹ These investigators also reported that the effect of bowel obstruction was influenced by the location of the tumor. The occurrence of bowel obstruction in the right colon was associated with a significantly diminished disease-free survival, whereas obstruction in the left colon demonstrated no such effect.²⁵⁹ Because tumors in the right colon can grow to a larger size, and therefore be present longer before causing symptoms of obstruction, it has been postulated that it is because of advanced stage at presentation rather than mechanical obstruction that these patients have a worse prognosis.²⁶⁰ After multivariate analysis, the Gastrointestinal Tumor Study Group (GITSG) concluded that obstruction was an independent prognostic factor for survival independent of Dukes stage.²⁶¹

Perforation is another untoward sign in patients with colorectal carcinoma. The GITSG reported that perforation was a prognostic factor for disease-free survival.²⁷⁹ In the Massachusetts General Hospital series, the actuarial 5-year survival for patients with obstructing and perforating carcinomas was 31% and 44%, respectively, whereas it was 59% in the control group.²⁶²

Perforation and obstruction cause an increase in morbidity and mortality. Both of these presentations can lead to electrolyte imbalance, dehydration, and infection, which add on to a surgical procedure, performed under emergent conditions. In general, patients presenting with an obstructing cancer should undergo resection if possible.

Two or three staged procedures are most commonly used for obstructing cancers. In the former, resection of the obstructing lesion is performed with a proximal ostomy. In the latter, a diverting stoma is performed as the first stage, followed by resection as the second stage, and takedown of the stoma as the third stage. The procedure of choice must be individualized depending on the condition of the patient, experience of the surgeon, and comorbid conditions. One-stage procedures can be performed in certain circumstances such as an obstructing right-sided colon carcinoma where a right hemicolectomy with primary anastomosis can be safely performed, or an obstructing distal transverse or splenic flexure carcinoma where consideration can be given to an extended right hemicolectomy or a subtotal colectomy. Patients with obstructing descending colon cancer could be considered for a subtotal colectomy.

One of the major concerns in the resection of obstructing lesions has been the bowel preparation. Several investigators have attempted to perform on table lavage so as to clean the bowel and perform a primary anastomosis. Others have used laser therapy to open the obstruction and pass a tube to decompress and clean the bowel or balloon dilatation and endoscopic stent placement so as to be able to decompress the patient and perform adequate bowel preparation followed by surgical resection. Only experienced individuals should only perform the latter procedures and only if the patient is in stable condition.

Perforation of the colorectal cancer occurs most commonly at the site of the tumor itself or in the bowel proximal to the obstruction. These patients can present with localized peritonitis or diffuse peritonitis. In addition, they can be dehydrated, septic, and with electrolyte imbalances. All these factors add to the morbidity of emergent surgery. Perforation can result in adherence to adjacent organs or fistula formation. Concomitant perforation is not uncommon in patients with obstruction.²⁶³ The principles of surgical management are the same as with patients with obstructing carcinomas. Ideally, an attempt at resection should be performed. However, there are times that the inflammatory reaction or patient's condition preclude resection, and thus, it is safer to divert the patient and drain the perforation rather than embark on an unsafe surgical procedure. A downside of this approach is seeding of the drainage tract. The tract must be removed en bloc with the tumor at the time of definitive resection. This potential downside must be weighed with the complications of an unsafe resection through unclear planes in a potentially unstable patient.

Contiguous Organ Involvement

It is not uncommon to encounter colorectal tumors adhered to adjacent organs. This contiguous involvement may be the result of bulky local growth by the tumor or by adhesions secondary to a local perforation and/or fistula formation. Any intraabdominal structure is at risk for adherence. Most commonly in colon cancer adhesions occur to the abdominal wall, to the duodenum, stomach, small bowel, ureters, urinary bladder, uterus, and ovaries. Rectal cancers most commonly invade the uterus and vagina, urinary bladder, and sacrum or coccyx. Surgery for these tumors must be carefully planned so as to resect the tumors en bloc without disturbing the adhesions. These adhesions are malignant in nature in over 40% of the cases. If en bloc resection is not performed in these patients, recurrence and survival will be compromised.²⁵¹ It is important to note that some colorectal tumors can attain large sizes and invade adjacent organs without lymph node metastases.²⁶⁴ Even in patients with stage III disease, if an en bloc resection is performed 5-year survival can approach 25%.²⁵¹

Peritoneal Disease

Peritoneal carcinomatosis carries a dire prognosis. Some patients will present with carcinomatosis at the time of diagnosis. In extremely highly selected situations cytoreductive surgery combined with intraperitoneal chemotherapy may benefit some of these patients.²⁶⁵

Hepatic Resection

The liver is the most common site of metastatic disease from colorectal cancer. It is estimated that 8% to 10% of all patients undergoing curative resection of colorectal primary tumors will have isolated metastases as the first sign of disseminated disease and that these are resectable in approximately 25% of the patients.²⁶⁶ Preoperative positron emission tomography (PET) scanning has been reported to alter the management in approximately 29% of patients by identifying areas of metastases not visualized by using conventional imaging.²⁶⁷ At the time of surgical exploration, intraoperative ultrasound should be undertaken because it demonstrates unsuspected additional lesions in approximately 5% to 6% of patients.²⁶⁸ A number of selection criteria define the likelihood of successful hepatic resection: primary tumor stage II versus stage III, B versus C, size of metastases, number of metastases, satellitosis, time to recurrence, CEA level, extent of margin, and extent of surgery.²⁶⁹ The 5- and 10-year survival from hepatic resection is about 30% and 20%, respectively.²⁶⁶ Hepatic arterial infusion (HAI) of floxuridine (FUDR) with or without continuous infusion of 5-FU after potentially curative resection of hepatic metastases was evaluated in three prospective studies.^270–272 The German Cooperative study was terminated after an interim analysis because there was no improvement in survival in patients treated with HAI with 5-FU and folinic acid versus those treated with surgery alone.²⁷⁰ The other multicenter study conducted in the United States, revealed that HAI with FUDR combined with systemic 5-FU and folinic acid reduced recurrence rate but did not improve overall survival when compared to surgery alone.²⁷¹ The third trial, a single-institution trial that evaluated HAI plus systemic 5-FU and folinic acid versus systemic 5-FU and folinic acid without HAI revealed a decrease in hepatic recurrence rate and improvement in actuarial overall survival at 2 years.²⁷² It is important to note that a fair number of patients in these trials could not receive the total dose of chemotherapy assigned (HAI). In selected patients, re-resection of isolated hepatic metastases is an option and has provided encouraging results.²⁷³ In extremely highly selected patients, combined resection of hepatic and pulmonary metastases has been performed.²⁷⁴

When metastases in the liver are widespread or nonresectable because of size and/or anatomic location, different locoregional modalities of treatment, such as intraarterial hepatic chemotherapy by catheter or via implanted pump, ethanol intratumoral injections, radiofrequency ablation, and cryoablation, have been investigated.^275–278 Each of these approaches remains investigational. Prospective randomized trials and further followup are necessary to determine their objective value.

Systemic chemotherapy has been used to decrease tumor size and to convert unresectable lesions to resectable lesions.²⁷⁹ An added benefit of this approach is that it enables the clinician to observe the natural history of the tumor while the patient is being treated. When an extended resection is contemplated, selective portal vein embolization to hypertrophy the remaining segment of the liver has been performed.

Laparoscopic Colectomy

Laparoscopic-assisted colectomy is a minimally invasive technique that has been gaining popularity in the management of colon and rectal cancer. The procedure consists of mobilizing the colon laparoscopically and performing an extracorporal anastomosis. In some situations, especially in the low sigmoid colon, an intracorporal anastomosis can be performed. In rectal cancer the technique is used to mobilize the mesorectum. The main potential benefits from laparoscopic colectomy are earlier postoperative recovery and less narcotic use. In earlier trials, there was no difference in outcome, staging, number of lymph nodes resected, and survival.^280,281 In a prospective randomized single-institution study comparing open versus laparoscopic colectomy, at a median followup of 43 months, Lacy and colleagues reported an improved cancer-related survival in patients undergoing laparoscopic surgery.²⁸² The cancer-related improvement in survival was apparently the result of an improvement in laparoscopic surgery in stage III patients. Morbidity and hospital stay were improved in the laparoscopic surgery group.²⁸² In the United States, results from the Intergroup prospective randomized trial have not been published. However, a short-term quality-of-life outcome study based on this trial was reported.²⁸³ Minimal short-term quality-of-life benefits were found in the laparoscopic-assisted resection patients as compared to the open resection group.²⁸³ These authors concluded that until the ongoing laparoscopic versus open colectomy trial establishes the effectiveness of the former in preventing recurrences and death from colon cancer, laparoscopic-assisted procedures should not be offered to patients with potentially resectable disease.

Distal Margin

Proximal and distal margins of resection must be adequate. A distal surgical margin of at least 2 cm in the fresh specimen is desired. Rarely, rectal adenocarcinomas will have distal submucosal spread. In patients with distal rectal adenocarcinomas at least a 1-cm margin of resection is desirable. Some authors report margins of less than 1 cm in patients who received neoadjuvant chemoradiation and underwent sphincter-saving procedures without compromising either recurrence or disease-free survival.^285,286

Lymphovascular Ligation

Lymphovascular ligation in rectal cancer resection should be performed at the origin of the superior rectal artery. There is no evidence that high ligation of the IMA offers any benefit over ligation at the level of the origin of the superior rectal artery.

Mesorectal Excision

Local recurrence after surgery alone for operable rectal adenocarcinoma is a major problem. Its incidence after surgery alone is reported to be between 25% and 50% for stages II and III rectal adenocarcinoma.²⁸⁷ These high recurrence rates are probably the result of incomplete dissection of the mesorectum. TME consists of sharp dissection of the plane between the endopelvic fascia and the mesorectum, with removal of the mesorectum with its intact fascia propria and with preservation of the pelvic fascia and the autonomic nerve plexus. The mesorectum addresses distal extramural spread and circumferential margins of excision. In the Dutch Mesorectal Trial, a randomized prospective trial, local recurrence was 8.2% after TME alone for resectable rectal cancer.²⁸⁴ Discontiguous spread in the mesorectum varies from 5% to 64%.²⁸⁸ Lymph nodes account for the majority of the discontiguous spread. Circumferential or radial margin involvement by tumor is a prognostic factor for both local and distant disease. The radial margin of excision can also serve as an indicator of the quality of the surgery.²⁸⁹ In the Dutch Mesorectal Trial, in a sample of 656 patients who underwent TME alone, 18.3% (n = 120) had a positive circumferential margin after curative resection. Patients undergoing abdominoperineal resection (APR) had a higher incidence (28.8%) than did those undergoing low anterior resection (LAR) (13.5%).²⁹⁰ Sixteen percent of patients with positive margin developed local recurrence within 2 years of followup.²⁹⁰ Patients who had a negative circumferential margin of between 1 and 2 mm had a local recurrence rate similar to that experienced by those with positive margins (14.9% vs 16.4%), whereas the local recurrence rate in patients with margins greater than 2 mm was between 2.4% and 10.3%.²⁹⁰ Distant metastases occurred in 37.6% of patients with positive circumferential margins, whereas it was 12.7% in patients with > 1 cm margin.²⁹⁰ Systematic macroscopic evaluation of the mesorectum was performed in 180 specimens from patients undergoing TME.²⁹¹ Almost 24% of the patients had an incomplete mesorectal excision as defined by the pathologist.²⁹¹ Even though the followup was short, there was a statistically significant difference in survival in patients with a macroscopically incomplete TME.²⁹¹ However, there was no statistically significant difference in local recurrence.²⁹¹

Extended Lateral Pelvic Node Dissection

Routine internal iliac node dissection is not recommended in rectal cancer surgery. Metastases to these lymph nodes occur in approximately 10% of patients and imply a dire prognosis. If clinically indicated, an attempt to remove these lymph nodes at the time of surgery is warranted.

Inadvertent Perforation

Retrospective studies demonstrate that inadvertent perforation at the time of surgery statistically reduces 5-year survival and increases local recurrence.²⁷² The incidence of inadvertent perforation during rectal cancer surgery is reported to be between 7% and 25%.²⁵⁴

Local Therapy for Rectal Adenocarcinoma

Local therapy has been evaluated in selected patients with rectal adenocarcinoma. These therapies include electrofulguration, endocavitary radiation with or without brachytherapy, and local excision with or without external beam radiation or chemoradiation.^292–295 Of these techniques, local excision has the advantage of providing a specimen for pathologic evaluation. Surgical approaches to local excision include transanal, transsacral, or transsphincteric approaches. The latter two procedures are less-commonly performed than is the transanal procedure.

Full-thickness transanal local excision has been used for selected small (< 4 cm) rectal tumors. Patients are staged clinically with history and physical examination, computed tomography (CAT) scan of the abdomen and pelvis, and transrectal ultrasound. Patients with clinically enlarged lymph nodes or fixed tumors are not considered good candidates for local excision. Ideally, the tumor should be easily palpated during rectal exam, be mobile, and should not involve greater than 40% of the bowel circumference. After full-thickness excision, the histopathology is examined. Patients without disease extension to the muscularis propria (T1 tumors) and with favorable histology (moderate or well differentiated) have a local failure risk of < 10%. Because of the low overall risk of recurrence, postoperative chemoradiation is not usually recommended. Patients who have poorly differentiated T1 tumors, or who have T2 tumors with any differentiation, have routinely been treated with postoperative chemoradiation to reduce the risk of pelvic recurrence. Initial reports with short followup were very encouraging and stimulated a significant amount of enthusiasm. Postoperative radiation is recommended among T1 tumors with unfavorable characteristics or T2 tumors because the local disease recurrence risk is about 20%.^296,297

In the CALGB multiinstitutional trial of local excision for T1 tumors and local excision and chemoradiation for T2 tumors, at a median 48 months followup, the 6-year actuarial survival and failure-free rates for the eligible T1 and T2 patients were 85% and 78%, respectively.²⁹⁴ There were 4 failures (2 local only, 1 distant only, and 1 local and distant) in 59 T1 patients. In 51 T2 patients, there were 10 recurrences (5 local only, 2 local and distant, and 3 distant only). The salvage rate for local recurrences only was greater than 50%, but with a short followup reported.²⁹⁴ A retrospective analysis from the University of Minnesota revealed a recurrence rate of 18% and 37% in patients undergoing local excision alone for T1 and T2 tumors respectively.²⁹⁵ These investigators subsequently published their results of salvage radical surgery in 29 patients after failed local excision.²⁹⁸ Twenty-three of 29 patients had surgery with curative intent. At a mean followup of 39 months, the disease-free survival was 59%.²⁹⁸

Local excision has been used after preoperative (chemo) radiation in medically inoperable patients and in patients who refuse APR.^299–301 tumors after neoadjuvant chemoradiation has been reported.^300,301 However, the followup in the latter studies is short, and there are reports of lymph node metastases in specimens of pathologically T0 tumors after neoadjuvant chemoradiation.³⁰² Local excision after neoadjuvant therapy is investigational and should be only performed in the context of a clinical trial.

Endocavitary radiation (contact therapy) is a technique that administers very-high doses of radiation. Each dose is concentrated in the tumor with minimal radiation dose to surrounding tissues. This treatment is appropriate for medically inoperable patients and for patients who refuse surgery. The technique was pioneered and developed in Europe where it is still commonly used with results comparable to the results with transanal excision, endoscopic mucosal resection, and fulguration for T1 tumors. Patients with T2 and T3 tumors are best treated with combinations of pelvic chemoradiation, the application of radioactive sources (brachytherapy), and contact therapy to address the risk of lymph node involvement and extramural extension.³⁰³ Currently in the United States, this technique is only performed in a few centers.

Pelvic Recurrence

Because of the use of radiation, chemoradiation, and TME, pelvic recurrences after curative surgery for rectal adenocarcinoma have decreased. Still, there is a significant number of patients who present with isolated pelvic or perineal disease in whom re-resection may be warranted. Surgical approaches for recurrent disease include re-resection with a low anterior resection, abdominoperineal resection, pelvic exenterations (or its modifications), and, occasionally, extended sacral resections. If chemoradiation was not used during the management of the primary tumor, preoperative chemoradiation may be warranted prior to re-resection. It must be mentioned that these procedures are generally performed in conjunction with urology, neurosurgery, and plastic surgery. If available, intraoperative radiotherapy should be strongly considered. After re-resection of isolated pelvic recurrences, there is a 20% 5-year survival. It cannot be overemphasized that these procedures carry a significant morbidity and mortality and should be performed by experienced surgeons. Other forms of palliation should be considered prior to undertaking an aggressive surgical approach in a patient in whom cure will not be attainable.

Surgical Prognosis

With improvements in anesthesia and perioperative management, operative morbidity and mortality have been reduced. Therefore, extensive procedures are being performed with better results than in the past. In the 1990s, for the first time in colorectal cancer, adjuvant therapy was proved to be effective in decreasing recurrence and improving survival. The new decade will hopefully bring even better adjuvant therapy with newer radiation techniques and molecular targeting of tumors.

Surgical cure of colorectal cancer is determined by stage of the tumor and its biologic behavior. Early colorectal cancers (T1-2N0M0) can be cured with surgery alone. The 5-year survival for patients with stage II colorectal cancer is approximately 70%. Surgery alone in stages II and III rectal cancer carries a high failure rate; postoperative chemoradiation is warranted if the patients were not treated with neoadjuvant chemoradiation. Because the 5-year survival in stage II (T3N0M0) colon cancer patients is approximately 77% with surgery alone, adjuvant therapy is not widely recommended for these patients. As previously discussed, molecular markers may help in deciding adjuvant therapy for a subgroup of these patients. T4N0M0 colon cancer patients are at high risk of both local and distant recurrence and should be considered for adjuvant chemoradiation.³⁰⁴ Stage III colorectal cancer patients (any T N1-2 M0) have a 20% to 50% 5-year survival rate after surgery alone. It is in stage III colon cancer patients that adjuvant therapy has proven to be effective.

Followup After Primary Therapy for Colorectal Cancer

Sixty percent to 80% of recurrences after curative surgery for colorectal cancer occur within the first 2-years of therapy. Followup should be more intensive during this time. There is a wide variation in followup practices after curative resection. Some of these practices include intensive followup, while other practices are less intensive in the followup. The American Society of Clinical Oncology has published surveillance guidelines after colorectal cancer resection.¹⁸¹

Intergroup Postoperative Rectal Trials

After the role of concurrent chemoradiation was established, subsequent trials were conducted by the Gastrointestinal Intergroup that evaluated the schedule of administration and biomodulation of 5-FU. A Phase III trial conducted by the Gastrointestinal Intergroup compared protracted venous infusion 5-FU to bolus 5-FU with concurrent radiation therapy in a trial that randomized 660 patients with resected stages II and III rectal cancer. The value of the addition of methyl-cyclohexylchloroethylnitrosurea (CCNU) was also tested by using 2 × 2 randomization. Adjuvant bolus 5-FU (500 mg/m²) was given for two cycles prior and two cycles following chemoradiation in both arms. Patients received either bolus 5-FU (500 mg/m²) for 3 consecutive days during weeks 1 and 5 of radiation therapy, or protracted venous infusion 5-FU given at a rate of 225 mg/m²/d, 7 days per week during pelvic radiation. This trial showed significant improvements in relapse-free and overall survival in the infusional arm. Local control rates were similar. The toxicity profiles were also different; bolus 5-FU had more hematologic toxicity and infusional 5-FU had more diarrhea. The addition of semustine to 5-FU before and after chemoradiation was found to be of no benefit. A meta-analysis was also reported that compared the efficacy of 5-FU in 1,219 patients with advanced colorectal cancer who received 5-FU either by bolus or continuous infusion on clinical trials. Tumor response and overall survival rates were significantly higher in patients who had received protracted venous infusion 5-FU. Hematologic toxicity occurred more frequently with bolus infusion (31% vs 4%; P < 10^-16), and hand-and-foot syndrome occurred more frequently with infusional 5-FU (34% vs 13%; 10^-7).³¹⁷

A subsequent trial (Int 0114) was conducted to evaluated the role of biomodulation. As in the previous Intergroup trial, patients received 2 months of systemic chemotherapy before and after concurrent chemoradiation. In a four-arm study, 1,696 patients with resected rectal cancer were randomized to receive four different chemotherapy regimens: bolus 5-FU alone, 5-FU with leucovorin, 5-FU with levamisole, or 5-FU with leucovorin and levamisole. The trial was designed before the results of the previous Intergroup study were known, so protracted venous infusion 5-FU was not used with radiotherapy. Instead, bolus 5-FU with or without leucovorin was used with radiotherapy in all four arms. There was no survival difference between the arms, after a median followup duration of 48 months, but there was a significantly greater acute gastrointestinal toxicity rate in the three-drug arm as compared to bolus 5-FU alone.³¹⁸

The subsequent Intergroup study (Int 0144) has reached its accrual goal but data are not available from it yet. This study was designed to compare the current standard postoperative therapy for resected rectal cancer (4 months of 5-FU and leucovorin given on the Mayo Clinic schedule) and pelvic radiotherapy with concurrent protracted venous infusion 5-FU (225 mg/m²/d, 7 days per week) to two experimental arms incorporating biomodulation and protracted venous infusion throughout, respectively. Radiotherapy is traditionally given after 2 months of chemotherapy as in the previous two Intergroup studies, but this practice is empiric and may not be the best way to sequence therapy. In fact, there are randomized data indicating that delivering radiotherapy during the first 2 months of therapy may have a disease-free survival advantage.³¹⁹ It is probably best to assess the relative distant and pelvic failure risks and administer chemotherapy first if the distant risk is more significant (ie, multiple positive nodes) and pelvic radiotherapy if the local therapy risk is more significant (ie, extensive transmural disease or narrow margin).

The NSABP Postoperative Rectal Trials

The NSABP R-01 trial is of historical interest because it was the only one of the early trials that did not include a combined modality arm, and it showed a survival benefit in males who received postoperative chemotherapy. Patients (n = 574) were stratified by age, stage, and sex, and then randomized to (1) surgery alone, (2) surgery followed by postoperative radiotherapy (46 to 47 Gy in 26 to 27 fractions), or (3) surgery followed by postoperative chemotherapy (semustine, vincristine, and 5-FU, or MOF). Only 86% of the patients received the total prescribed dose of radiation. There was a decrease in pelvic tumor recurrence in the postoperative radiotherapy arm (25% vs 16%; p = .06) but this did not lead to an improvement in overall survival. The adjuvant chemotherapy alone arm had significantly improved disease-free survival and overall survival compared to surgery alone. On subgroup analysis, men had improved survival and females actually had a lower overall survival with chemotherapy.³²⁰

The NSABP R-02 trial was designed to build on the available randomized data at the time: the GITSG study and the NSABP R-01 study. In the GITSG study, there were no significant differences in disease-free or overall survival between the chemotherapy alone and combined modality arms of the study, and in the NSABP R-01 study, MOF chemotherapy alone was beneficial to male patients and harmful to female patients. Therefore, the NSABP R-02 trial was designed to answer two questions in patients with resected stages II and III rectal cancer. First, is combined modality therapy superior to chemotherapy alone, and second, is MOF chemotherapy better than 5-FU and high-dose leucovorin (the Roswell Park regimen) in males? Male patients were subjected to a 2 × 2 randomization and it was recognized that the study could only answer the second question if there were large differences seen. As in previous trials, radiotherapy improved pelvic disease control (13% vs 8%; p= .02) but did not improve disease-free or overall survival. The concurrent bolus 5-FU that was used in this study has since been shown to be inferior to protracted venous infusion 5-FU,³²¹ which could partly explain the lack of a survival difference. In the global test for interactions of covariates that were significant in the multivariate model that was used, subgroups of patients did appear to have a disease-free and/or overall survival benefit from radiotherapy (patients < 60 and those undergoing APR).³¹² This finding underscores the notion that patients can probably be selected who have a low pelvic recurrence (and survival) benefit from pelvic radiotherapy. The pelvic recurrence rates in the surgical control arm in this study were lower than in any other randomized trial. Of note is that pelvic imaging was not done in a systematic fashion in followup, only first sites of failure were recorded, and diagnostic imaging was not reviewed centrally. Therefore, it is probable that pelvic failure was underestimated in both arms. In spite of its limitations, this is the only randomized trial with enough power to compare chemotherapy alone to combined chemotherapy and radiotherapy in this population, and it can be used to support the use of chemotherapy alone in selected patients with resected stages II and III rectal cancer.

Preoperative versus Postoperative Adjuvant Chemoradiation

There are several significant advantages of preoperative chemoradiation over postoperative chemoradiation in rectal cancer treatment. Mobilizing small bowel out of the radiotherapy field can reduce toxicity. This can be accomplished more commonly in patients who have not undergone laparotomy.³²² The use of preoperative chemoradiation can lead to increased sphincter preservation rates,^323–329 and because the irradiated rectum is removed and unirradiated colon can be used to form a reconstructed rectum, long-term rectal function is probably better, although these points are controversial. Tissue perfusion is uninterrupted in patients who are treated preoperatively. This leads to better drug delivery and better oxygenation. Because oxygen enhances the effects of ionizing radiation, radiotherapy efficacy should increase. Randomized data also indicate that preoperative radiotherapy alone leads to increased local control³³⁰ than postoperative chemoradiation. The main disadvantage of preoperative chemoradiation is that because clinical staging has limitations, some patients may be treated unnecessarily. The number of such patients can be minimized by using endoscopic ultrasound. When surgery is performed first, patients can be selected for postoperative chemoradiation, based on T-stage, N-stage, margin status, distance from the anal verge, and other prognostic factors. Some clinicians are concerned that the use of preoperative radiotherapy will increase perioperative morbidity. However, it appears that perioperative morbidity is not increased in centers that routinely use preoperative therapy. Surgical experience is an important factor in lowering perioperative morbidity rates in irradiated rectal cancer patients. For example, perioperative complication rates at M. D. Anderson decreased with time after the introduction of preoperative therapy.³³¹ The use of diverting ileostomy while the anastomosis heals is important to reduce perioperative anastomotic leaks, particularly in patients with low rectal anastomosis.

Prospective trials have been designed to address this controversy by the Radiation Therapy Oncology Group (RTOG) and NSABP, but the established practice patterns were so rigid that the studies suffered from poor accrual and had to be closed prematurely. The German CAO/ARO/AIO trial comparing preoperative chemoradiation to postoperative chemoradiation recently met its accrual goal of 800 patients.³³² It is hoped that this study will provide an answer to the question. The only randomized trial that has been completed comparing preoperative to postoperative radiotherapy was conducted in Sweden. Patients received either short-course preparative radiotherapy (25.5 Gy in 5 fractions in 1 week) or split course postoperative therapy (60Gy in 30 fractions in 8 weeks). There was pelvic control benefit, but no survival benefit,³³³ but the nonstandard fractionation and lack of the use of chemotherapy in both arms make this study difficult to interpret in the context of the current standard treatments in the United States.

Preoperative (Chemo)radiation

Currently, there are different standards of care in Europe than in the United States with regard to the dose and schedule of preoperative radiotherapy. In Europe, investigators have established the benefit of preoperative short-course radiotherapy (25Gy in 5 fractions in 1 week) alone followed by immediate surgery. In the United States, standard preoperative doses range from 45 Gy to 54 Gy and concurrent protracted venous 5-FU chemotherapy is routinely administered. This approach has been extrapolated from postoperative trials and is becoming more common.

Many trials conducted in Europe demonstrate a pelvic control benefit with the addition of hypofractionated radiotherapy over surgery alone, but only the Swedish Rectal Trial demonstrated a survival benefit. Preoperative radiotherapy (25 Gy in 5 fractions in 1 week) was followed by surgery within 1 week. Improvements were seen in local control (73% vs 89%; p < .001), overall survival (58% vs 48%; p = .004).³³⁴ This study was the only trial with enough power (1,168 patients) to detect small differences in survival. In addition to being underpowered, previous trials had high perioperative mortality rates in irradiated patients.³³⁵ The perioperative mortality rate has decreased with the use of smaller fields and multiple field techniques.³³⁶ A meta-analysis of 14 randomized trials evaluating preoperative radiotherapy (most using short course) versus surgical resection alone (6,426 patients) was performed. No overall increase in postoperative mortality was reported with radiotherapy except in the trials that failed to use a multiple field technique. Preoperative radiotherapy was found to reduce the 5-year overall mortality rate, cancer-related mortality rate, and local recurrence rate.³³⁷ Furthermore, a detailed analysis by Suwinski and colleagues looked at the effect of time, dose, and fractionation factors by using published data from both single arm and comparative trials, as well as from retrospective studies reporting results using preoperative radiotherapy alone. The authors concluded that the percent reduction in pelvic recurrence is comparable in the accelerated hypofractionated schedule (25 Gy in 5 fractions in 1 week) compared to the standard schedule (45 to 50 Gy at 2 Gy/fraction).³³⁸ Thus, from an efficacy standpoint, the approaches are probably similar.

However, the potential disadvantages of the short-course approach are that time is not usually allowed for tumor regression prior to surgery, there may be increased perioperative and/or late morbidity and mortality, and the radiosensitizing effects of concurrent chemotherapy cannot be fully exploited. Tumor regression with time may contribute to sphincter preservation.³²⁹ This question was the topic of a randomized trial conducted in Lyon, France. Patients were treated with preoperative radiotherapy alone (39 Gy in 13 fractions) and randomized to a 2-week surgical interval or a 6- to 8-week interval. Pathologic response was significantly increased after the longer interval (26% vs 11%).³²⁴ With regard to toxicity of short-course therapy, the Stockholm trials identified advantages of combined modality therapy, but significant late morbidity and mortality was also observed. Neuropathy, thromboembolism, femoral neck and pelvic fractures, intestinal obstruction, postoperative fistulae, and perioperative death³³⁶ all occurred with a higher frequency than one would expect with standard fractionation. To a certain extent, these events were reduced with improved techniques (multiple fields, and dropping the superior border from L2 to the L5/S1 interspace)^334,339 and the most recent studies of hypofractionated preoperative radiation (the Swedish Rectal Trial and the Dutch Colorectal Cancer Group Trial)^334,340 reported lower treatment-related morbidity and mortality. In Europe, the use of concurrent chemotherapy with radiotherapy in the preoperative setting is controversial. In the United States, its use in the preoperative setting is extrapolated from postoperative data.³⁰⁸

Preoperative Radiotherapy in the United States

The majority of institutions that use preoperative chemoradiation in the United States today use 45 to 54 Gy and concurrent protracted venous infusion 5-FU (225 to 300 mg/m²/d). This approach has been adopted without randomized trials based on the positive results of concurrent chemoradiation in the GITSG and NCCTG trials.^306,308 There are no completed randomized comparisons between preoperative radiotherapy alone and preoperative chemoradiation. Investigators at the University of Florida were among the first in the United States to adopt preoperative radiotherapy. About the time that the positive results of the NCCTG postoperative trial were reported, a relatively abrupt policy change took place and concurrent 5-FU was administered with the preoperative radiotherapy. Comparing the early (1975 to 1990; n = 219) to the late era (1991 to 1997; n = 109), an overall survival difference was seen in the late era, when concurrent chemotherapy was used (87% vs 58%; p = .0005).³⁴¹ Many other things changed between 1975 and 1997, including surgical techniques, radiotherapy techniques, supportive care, and diagnostic imaging. However it appeared from their analysis that there was not a gradual improvement among the suberas, indicating that these potential confounding factors may not have played a significant role in the improved outcome.

Adjuvant Chemotherapy in Addition to Chemoradiation and Surgery

The standard postoperative adjuvant chemoradiation regimen includes 4 months of 5-FU-based systemic therapy in addition to pelvic chemoradiotherapy. This standard evolved empirically from pattern of failure observations from the four-arm Phase III trial conducted by the GITSG that showed a reduction in the rate of distant metastases with chemotherapy alone and a reduction in pelvic recurrence with radiotherapy alone.³⁰⁶ Subsequent Gastrointestinal Intergroup trials have included systemic chemotherapy before and after pelvic chemoradiation.^318,342 However, there is not uniform agreement regarding the role of chemotherapy in addition to preoperative chemoradiation. Because most patients have pathologically negative nodes following preoperative chemoradiation,³⁴³ some clinicians have argued that systemic therapy is not indicated. This argument is partly due to the lack of a proven survival benefit of chemotherapy in node negative colon cancer. The controversy is illustrated by the fact that the European Organization for the Research and Treatment of Cancer (EORTC) is conducting a Phase III trial in which patients are randomized to receive either 5-FU-based chemotherapy or no further therapy following preoperative radiation and surgery. It is hoped that this study will answer that question, but until these data are available, the use of adjuvant 5-FU-based chemotherapy can be supported on the basis of the postoperative trials. Although it appears in some series that patients who have an excellent pathologic response to preoperative chemoradiation have a more favorable outcome,^344,345 it is unknown whether these patients benefit from further chemotherapy. Because of their lower overall absolute risk, there is likely a potential smaller absolute benefit.

Radiotherapy Technique and Volume of Irradiation

Bowel exclusion techniques can be used to move the bowel out of the irradiated pelvis. Many institutions use a three-field technique with an open tabletop (“belly board”). This technique has been shown to reduce the incidence of bowel obstruction in an analysis of 224 patients treated with postoperative radiotherapy at M. D. Anderson Cancer Center between 1973 and 1990.³⁴⁶ The median volume of small bowel was also reduced by 54% using a similar technique in patients receiving postoperative radiotherapy.³⁴⁷ Prone positioning with a three- or four-field technique without an open tabletop device also results in excellent bowel exclusion from the radiation field. A distal and anterior margin of 3 cm on the primary tumor should be used for low-lying lesions, and care taken not to irradiate the anal canal in higher lesions. In most cases, the anal canal is partially spared in the lateral fields. The internal iliac and the mesorectal lymph nodes are treated and the entire sacrum and coccyx are included in the treatment volume. The risk for inguinal node recurrence is low (< 5%) if the primary tumor extends to the anal canal and the inguinal lymph nodes are not treated. Treatment of the inguinal nodes increases the volume of irradiation and along with it the risk of severe skin reactions, scrotal edema, and diarrhea. For these reasons, elective inguinal nodal irradiation should not be routinely given.³⁴⁸ However, the inguinal region should be treated if inguinal lymph nodes are clinically enlarged. Cytologic confirmation of inguinal nodal involvement is recommended. Metastatic involvement in the inguinal nodes portends an especially poor prognosis for dissemination of disease.^349,350 Similarly, the risk of external iliac nodal recurrence is low without elective irradiation in patients with T4 tumors.³⁵¹

Management of T4 and Recurrent Rectal Cancer

Criteria to establish T4 disease clinically are inherently subjective. As such, inclusion criteria for the groups of patients that have been reported as having T4, advanced, and “unresectable” disease have not been consistency applied. Studies have ranged from the use of no objective criteria for advanced disease^352,353 to the use of endoscopic ultrasonography and CT or MRI³⁵⁴ to increase the objectivity of clinical staging. These differences in staging evaluation and criteria for T4 disease make the comparison of these series difficult. The most objective way to classify T4 disease is to use abdominopelvic CT scan or MRI, endoscopic ultrasound, and pelvic examination to assess patients for evidence of adjacent organ or bony invasion. In patients with T4 disease treated with preoperative chemoradiation (median 45 Gy) followed by surgical resection at M. D. Anderson, the pelvic disease recurrence rate was 20%, the distant metastases rate was 44%, and the overall survival rate was only 50%.³⁵¹ Similar results have been reported with the use of postoperative chemoradiation after surgical resection for pathologically staged T4 disease.³⁵⁵

Recurrent rectal cancer can vary significantly in its clinical presentation. When local tumor recurrence is identified early and is isolated, curative surgical resection is the mainstay of treatment. If synchronous metastases are present or the tumor is unresectable, then palliative radiotherapy is usually preferred. Preoperative chemoradiation to 50.4 Gy, followed by resection and intraoperative radiotherapy (IORT, 10 to 20Gy) is the preferred approach. The goals are to increase resectability and pelvic control. A large experiment at the Mayo Clinic indicates that even with an aggressive approach involving preoperative radiotherapy and IORT, the subsequent pelvic recurrence rate in the radiotherapy field was 37% and the distant metastases rate was 54%. The 5-year overall survival was 20%, as compared to the 5% expected with surgery-alone historic controls.³⁵⁶ Even with the use of pelvic chemoradiation, aggressive surgery and adjuvant chemotherapy in T4 and locally recurrent rectal cancers, pelvic and distant disease recurrence are significant limitations of therapy. Future investigation of novel radiosensitizers and the development of systemic adjuvant regimens are needed to improve the outcome of T4 and recurrent rectal cancer.

Palliative pelvic radiotherapy can delay progression of recurrent local disease, but 5-year control is expected in less than 10%. Investigators from The Princess Margaret Hospital reported a study evaluating 519 patients with locally recurrent rectal cancer who had not been previously irradiated. The median survival time was 14 months after palliative irradiation alone and the median time to local disease progression was 5 months. The pelvic disease progression-free rate was 7% and the 5-year survival was 5%. Multivariate analysis showed that overall survival was correlated to the Eastern Cooperative Oncology Group (ECOG) performance status, absence of extrapelvic metastases, long intervals between the initial surgery and radiotherapy, total radiation dose, and absence of obstructive uropathy.³⁵⁷

Management of Patients Who Present with Synchronous Distant Metastases

The initial management strategy in patients who present with distant metastases from rectal cancer is controversial. Commonly, patients are treated initially with systemic therapy alone, with or without initial colostomy. This lack of attention to the primary disease can result in painful uncontrolled pelvic disease, which can be impossible to adequately palliate with radiotherapy or anything else. Additionally, many patients are unnecessarily left with a permanent colostomy. Initial chemoradiation can be used in patients who present with metastatic disease as well. Initial chemoradiation not only avoids nontherapeutic colostomy, but it also introduces a period of observation that can help in the selection of more aggressive subsequent therapy, thereby reserving resection of primary and metastatic disease for those who are most likely to benefit from it. Durable symptomatic palliation can be achieved in 80% of patients and colostomy can been avoided in 90%. The patients with progressive pelvic symptoms (usually obstructive symptoms) prior to death can be palliated with the use of endoscopic laser treatment, or colostomy at that time. After pelvic chemoradiation, patients can be selected for resection of their primary disease and other more aggressive therapies such as hepatic arterial infusion, radiofrequency ablation, and hepatic resection based on response to therapy, the extent and distribution of metastatic disease, and performance status. Patients with surgically incurable metastatic rectal cancer can be effectively palliated with initial hypofractionated chemoradiation (35Gy in 14 fractions in 3weeks, 30 Gy in 6 fractions in 3 weeks, or 36 Gy in 12 fractions in 2.5 weeks) and concurrent 5-FU (300 mg/m²/d, Monday to Friday).³⁵⁸ Decisions about further therapy can be made following restaging.